minutes from ietf53 ad-hoc meeting

[Florent Parent <]

Here's rough minutes from the discussion we had during lunch. Please,
corrent or fill out missing pieces.  Thanks for everyone who showed
up. We will need a room next time!

Participants (sorry for those I forgot the name, no blue sheet ;-)

George (APNIC), Joao Luis Silva Damas (RIPE), Shane Kerr (RIPE), David
Kessens, Cengiz, Larry Blunk, Susan Harris, ? (ARIN), Florent Parent,
Marc Blanchet, ? (RIPE), ? (RIPE).

Draft sent by Joao
- covers 2 method for aut-num
	1- define new attributes import6, import4, ...
	2- use generic import-policy + afi (similar to
draft-parent-multiprotocol-rpsl-00.txt)
- defines new classes for IPv6 (route6, route6-set, peering6-set,
filter6...)

draft-parent-multiprotocol-rpsl-00.txt
- Concerns raised on breaking current tools that access database with
  IPv6 objects, even if new tools would be required anyhow.

Proposition:
Converge the two approach. The following approach will be worked on:
- Try to minimize new IPv6 only classes
- Don't define IPv6 specific route-set, peering-set, *-set classes 
- Define generic class sets instead (name to be defined)
- use generic import-policy + afi 
- route6 is a special case. keep it.

- Aim for rapidly deployable solution
- document impact on servers (ex: mirroring)
- document impact on tools
- list pros and cons for each approach (duplicate classes for IPv6 vs.
direct IPv6 support)

TODOs:

Joao and Florent
- make sure that a new draft based on above proposition is out within for
next RIR meeting (in ~1 month)
- Aim to move draft forward for next IETF (July).

Larry Blunk
- Considering writing draft on new auth mechanism ?